Accuracy of spleen stiffness measurement in detection of esophageal varices in patients with chronic liver disease: systematic review and meta-analysis

2014 Clinical Gastroenterology & Hepatology 12;6 (935-945 e934)

BACKGROUND & AIMS: Spleen stiffness measurement (SSM) is a promising noninvasive alternative to esophagogastroduodenoscopy (EGD) that could be used in the diagnosis of esophageal varices (EV) in patients with cirrhosis. However, its overall diagnostic accuracy in various clinical settings is unknown. We conducted a systematic review and meta-analysis of studies that compared the accuracy of SSM with that of EGD in detecting EV in patients with chronic liver disease. METHODS: Through a systematic search of bibliographic databases and conference proceedings, and contact with authors, we identified 12 studies that reported the accuracy of SSM, compared with EGD, in the diagnosis of any and/or clinically significant EV in adults with chronic liver disease. In a meta-analysis, we combined measures of test performance of individual studies. RESULTS: Based on pooled estimates, SSM detected the presence of any EV with 78% sensitivity (95% confidence interval [CI], 75%-81%), 76% specificity (95% CI, 72%-79%), a positive likelihood ratio (LR) of 3.4 (95% CI, 2.3-4.9), a negative LR of 0.2 (95% CI, 0.1-0.4), and a diagnostic odds ratio of 19.3 (95% CI, 7.5-49.8). In a meta-analysis of 9 studies, SSM detected the presence of clinically significant EV with 81% sensitivity (95% CI, 76%-86%), 66% specificity (95% CI, 61%-69%), a positive LR of 2.5 (95% CI, 1.7-3.9), a negative LR of 0.2 (95% CI, 0.1-0.5), and a diagnostic odds ratio of 12.6 (95% CI, 5.5-28.7). There was significant heterogeneity among studies owing to differences in elastography techniques and study locations. The included studies that were at risk for spectrum bias, review bias, and disease progression bias. CONCLUSIONS: Based on a meta-analysis, current techniques for measuring spleen stiffness are limited in their accuracy of EV diagnosis; these limitations preclude widespread use in clinical practice at this time.

Pubmed : 24055985